A Canonical Particle Swarm Optimization(C-PSO) Approach to Identify High Utility Itemset

Abstract

High Utility Itemset Mining (HUIM) is the advanced process of identifying highly profitable items by assessing the unit profit of each item within extensive transaction databases. Over recent years, HUIM has emerged as a crucial subject with broad applications. It facilitates the identification of profitable items based on factors such as profit and quantity, setting it apart from conventional algorithms. Numerous algorithms have been developed to mine High Utility Itemsets (HUIs), addressing the challenge of searching for these sets in databases containing many distinct items. HUIM incorporates optimization techniques to reduce complexities and discover optimal solutions. This research uses Canonical Particle Swarm Optimization (C-PSO) to identify high-profit items, effectively managing convergence properties.Canonical Particle Swarm Optimization (C-PSO) is motivated by the imperative to enhance the efficiency of HUIM. C-PSO aims to swiftly and accurately identify high-profit items within transaction databases by optimizing the search space. Its advantages lie in superior convergence properties, addressing challenges in navigating extensive item spaces, and its adaptability to diverse utility mining scenarios, making it a robust and superior choice compared to existing techniques. The experimental results, conducted on three benchmark datasets, illustrate that C-PSO outperforms existing state-of-the-art techniques in the context of HUIM.